Tm JOURNAL OF B I O ~ I C A L CHE~~STRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol . 269, No. 47, Issue of November 25, PP. 29867-29873,1994 Printed in U.S.A.

Neutralizing Interaction between Heparins and Myotoxin 11, a Lysine 49 Phospholipase 4 from Bothrops asper Snake Venom IDENTIFICATION OF A HEPARIN-BINDING AND CYTOLYTIC TOXIN REGION BY THE USE OF SYNTHETIC PEPTIDES AND MOLECULAR MODELING*

(Received for publication, June 30, 1994, and in revised form, September 8, 1994)

Bruno Lomonte$Ofl, Ernest0 Morenoll**, Andrej TarkowskiS $$, Lars k HansonS, and Marco MaccaranaOO From the Departments of $Clinical Zmmunology, IWedical Biochemistry, and $$Rheumatology, University of Giiteborg, Goteborg, Sweden, the 5Znstituto Clodomiro Picado, Facultad de Microbiologia, Universidad de Costa Rica, Sun Jos$ Costa Rica, the **Center of Molecular Zmmunology, Havana, Cuba, and the #Department of Medical and Physiological Chemistry, Uppsala Biomedical Center, Uppsala University, Uppsala, Sweden

Heparin binds to phospholipase 4 myotoxins from Bothrops asper snake venom, inhibiting their toxic ac- tivities. This interaction was investigated using purified myotoxin 11, a Lys-49 phospholipase A, of this venom, and a series of heparin variants, fragments, and other glycosaminoglycans. The binding was correlated to toxin neutralization, using endothelial cells as a target. Myotoxin I1 binds radiolabeled heparin in solution un- selectively, and forms macromolecular complexes with an optimum at a heparin:toxin molar ratio of 1:5. Both O-sulfates and N-sulfates play a role in heparin binding, in the order of importance 2-O-sulfates > 6-Osulfates > N-sulfates. The shortest heparin oligosaccharides inter- acting with myotoxin I1 are hexasaccharides. The bind- ing of a neutralizing monoclonal antibody (MAb-3) to myotoxin I1 was not inhibited by heparin, indicating that the two molecules interact with different sites on the toxin. A synthetic peptide (residues 115-129 in the numbering system of Renetseder et al. (Renetseder, R., Brunie, S., Dijkstra, B. W., Drenth, J., and Sigler, P. B. (1985) J. BioZ. Chem. 260, 11627-11634) of myotoxin I1 displays both heparin-binding and cytolytic activities. It is concluded that heparin neutralizes myotoxin I1 by binding to a strongly cationic site in the region of resi- dues 115-129, a possible contribution of lysines 36 and 38 suggested by molecular modeling studies. As this cati- onic region appears to be responsible for the cytolytic activity of the toxin, the present report constitutes the first identification of a cytotoxic region on a phospho- lipase 4 myotoxin.

Phospholipases4 (P-S; EC 3.1.1.4)’ in snake venoms have acquired a variety of pharmacologicalhoxic activities, including neurotoxicity, myotoxicity, anticoagulant effect, and inflam-

* This study was supported by the Swedish Agency for Research Co- operation with Developing Countries (SAREC), Kabi-Pharmacia (Stock- holm, Sweden), Consejo Nacional para Investigaciones Cientificas y “ecnol6gicas de Costa Rica (CONICIT), International Foundation for Science (F/1388-2), Swedish Medical Research Council Grant 2309, and Polysackaridforskning AB (Uppsala, Sweden). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in ac- cordance with 18 U.S.C. Section 1734 solely to indicate this fact. ll To whom correspondence should be addressed. “el.: 506-229-0344;

Fax: 506-292-0485. The abbreviations used are: P-, phospholipase 4; GAG, glyco-

saminoglycan; HexA, unspecified hexuronic acid; GlcN, 2-amino-2-de- oxyglucose (D-glucosamine); IdoA, L-iduronic acid; GlcA, D-glucuronic acid; PBS, phosphate-buffered saline; r.m.s., root mean square.

matogenicity (1). Although their toxic effects have been a priori conceived as being consequence of the enzymatic activity, a growing body of evidence indicates that some toxic actions do not depend on phospholipolytic activity (1, 2). In the case of non-neurotoxic P- myotoxins, clear examples of the lack of requirement of an intrinsic enzymatic activity for the induction of muscle necrosis (3-7) or anticoagulant effect (8) have been presented. Nevertheless, the molecular site responsible for the toxic activity of P- myotoxins has remained elusive (1).

Heparin is a sulfated glycosaminoglycan (GAG) composed by alternating hexuronic (glucuronic (GlcA) or iduronic (IdoA) acid and glucosamine (GlcN) units. Sulfate substituents are mainly N-sulfate groups (at C-2 position of GlcN) or O-sulfate groups (at C-6 of GlcN and C-2 of IdoA) (9). A previous study showed that both standard heparin and heparin with low affinity for antithrombin bind to myotoxic P-s from the venom of Both- rops asper, inhibiting their in vitro cytolytic and i n vivo myo- toxic activities (10). Neutralization of myotoxic activity of the venoms from at least three other taxa containing P w s struc- turally related to those ofB. asper was also observed (10). Thus, heparin and its derivatives, especially those with little or no anticoagulant activity, might have potential as a complemen- tary treatment of envenomations caused by several species of snakes.

The purpose of the present work was (i) to investigate some basic aspects of this protein-GAG interaction, and (ii) to ap- proach the molecular mechanism of myotoxin neutralization, through the search for a heparin-binding site. For these pur- poses, we utilized purified myotoxin 11, a Lys-49 phospholipase 4 of B. asper (5,111, synthetic peptides, and a series of heparin variants, fragments, and other sulfated GAGS. Special interest was given to the correlation between the binding interaction and the neutralization of cytolytic activity. Use of synthetic peptides helped to identify a myotoxin region involved both in interaction with heparin, and in cytolytic action. Experimental data were complemented by molecular modeling studies of the proposed heparin-binding and cytolytic region.

EXPERIMENTAL PROCEDURES Glycosaminoglycans-Heparin from pig intestinal mucosa (stage 14,

Inolex Pharmaceutical Division) was purified as described elsewhere (12) and radiolabeled by ,H-acetylating free amino groups through treatment with labeled acetic anhydride (13). The product had a specific activity of -0.6 x lo6 dpdnmol, assuming a molecular weight of 15,000.

6-O-Desulfation along with N-desulfation of heparin was achieved by treatment with dimethyl sulfoxide, 10% water at 110 “C for 5 h (14), and was followed by either re-N-sulfation (15) or N-acetylation (16). Com- positional analysis of the re-N-sulfated product indicated non-O-sul- fated-HexA-GlcNSO, (39%) and IdoA(2-OS03)-GlcNS0, (46%) as the

29867

29868 Heparin-binding Cytolytic Site of Myotoxic Phospholipase A,

predominant disaccharide units. 2-0-Desulfated heparin was obtained under alkaline conditions (17) and contained 21% non-0-sulfated HexA- GlcNSO, and 70% IdoA-GlcNSO3(6-OSO,) as the predominant disac- charide units.

Selective N-desulfation was obtained by treatment with dimethyl sulfoxide:H,O 19:l at 50 "C (18). The sample, quantitatively retaining 0-sulfate groups, was completely N-acetylated.

0- and N-desulfated heparin was prepared by treatment with 10% methanol in dimethyl sulfoxide at 100 "C for 2 h (14). The product was completely re-N-sulfated or completely N-acetylated. In both cases the preparations still retained -0.3 0-sulfate group/disaccharide unit lo- cated exclusively in position C-2 of the iduronic acid.

Alow sulfated heparan sulfate (-0.6 sulfate group/disaccharide unit) from human aorta (19) was provided by W. Murphy (University of Mo- nash, Australia). Chondroitin sulfate (bovine cartilage) and dermatan

(20). sulfate (pig intestinal mucosa) were obtained as described previously

Oligosaccharides-Even-numbered heparin oligosaccharides were generated by partial deaminative cleavage of the polysaccharide with nitrous acid (pH 1.5; cleavage at N-sulfated GlcN units) (21), essentially as described elsewhere (22, 23), and the resulting 2,5-anhydro-~-man- nose residues were reduced with either NaBH, or NaB3H, (Amersham Corp.). The resulting labeled oligosaccharides had a specific activity of -0.6 x lo6 dpm 3Wnmol. Specific activities of heparin and heparin- oligosaccharides were expressed on the basis of hexuronic acid content, as measured by the carbazole reaction (24).

Purification of Myotoxin 11 and Preparation of Synthetic Peptide- Myotoxin I1 was purified from the venom of B. asper (5), appearing as a single band of 15,000 Da in SDS-polyacrylamide gel electrophoresis (25) and cathodic native polyacrylamide gel electrophoresis (26).

ETGKNPAKSYGAYGC), 60-71 (KKDRYSYSWKDK), and 105-117 Three peptides, corresponding to residues 1-26 (SLFELGKMILQ-

(KKYRYYLKPLCKK) of myotoxin 11, respectively, with native endings, were synthesized by Chiron Mimotopes (Victoria, Australia) using Fmoc (N-(9-fluorenyl)methoxycarbonyl) strategy (27). In the numbering sytem of Renetseder et al. (28), these peptides correspond to residues 1-27,69-80, and 115-129, respectively. Their purity was at least 96,75, and 96%, respectively, as assessed by high performance liquid chroma- tography on a LiCrosphere 100RP-18 column or by mass spectrometry. The sequences were based on data from Francis et al. (11).

Interaction between Saccharides and Myotoxin 11-Myotoxin I1 was incubated at room temperature for 2 h with the appropiate saccharides in 200 or 300 pl of 50 mM Tris-HC1, pH 7.4, 130 m~ NaCl (TBS) con- taining 0.5 mg/ml of bovine serum albumin. The protein, along with any bound GAGs, was recovered by quick passage of the mixtures through nitrocellulose filters (Sartorius, pore size 0.45 pm; 25-mm diameter) which had been placed onto a 10-well vacuum-assisted manifold filtra- tion apparatus. The filters were prewashed twice with 5 ml of TBS, before application of samples, which were immediately followed by an- other two washings with the same buffer; each washing step was com- pleted within 5 s. Protein-bound radioactivity was determined after submersion of the filters in 2 ml of 2 M NaCl for 30 min; the filters were withdrawn and the eluate was mixed with 2 ml of water and 12 ml of scintillation mixture (OptiPhase, Pharmacia Biotech Inc.) and counted in a Beckman LS 6000IC scintillation spectrometer. No residual radio- activity could be detected on the filters (for reference, see Maccarana et al. (29)). The variation between duplicates was generally <5%.

Cytotoxicity Assay-The cytotoxic activity of myotoxin 11, and the neutralizing effect of GAGs, were quantified on an endothelial cell line (tEnd) (30) as previously described (10). Cells were grown in 96-well plates until near confluence, in Iscove's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum, 2 m~ L-glutamine, 5 x M 2-mercaptoethanol, and 0.05 mg/ml gentamycin. At the time of the assay, culture medium was replaced with 150 pllwell of medium with 1% fetal calf serum, containing 10 pg of myotoxin 11, which had been previously incubated in the same medium (15 min, room temperature) with each GAG. The fetal calf serum was lowered to 1% to minimize the basal lactate dehydrogenase (EC 1.1.1.27) activity of the medium. After 3 h of incubation at 37 "C, 100 pl of supernatant were assayed for lactate dehydrogenase (kit no. 500, Sigma). For 100 and 0% cytotoxicity values, cells were incubated with 0.1% Triton X-100-containing medium or plain medium, respectively. Control samples of GAGs without toxin were included. The cytotoxic activity of peptides, dissolved either in Iscove's medium or phosphate-buffered saline (PBS; 0.12 M NaCI, 40 mM sodium phosphate, pH 7.2), was determined similarly.

Enzyme Immunoassay for Competition Binding of MAb-3-"Ab-3, a monoclonal IgG, antibody (31) neutralizing the myotoxic (32) and cyto- lytic (10) activities of myotoxin 11, was utilized to determine if heparin

0 z 3

8

0 5 10 1 5 20

PROTEIN OR PEPTIDE ADDED (w)

FIG. 1. Binding of SH-labeled heparin to myotoxin I1 and syn- thetic peptides. Approximately 10,000 dpm of ,H-labeled heparin (-90 m) was added to increasing amounts of myotoxin I1 (5 pg, for instance, corresponding to 1.7 p~), or peptides, in a final volume of 200 pl of incubation buffer. The amounts of 3H-labeled filter associated with myotoxin I1 or peptides were determined by the filtration procedure through nitrocellulose described under "Experimental Procedures." Blank values without proteidpeptides gave 80 dpm. (0) myotoxin 11; (0) peptide 115-129; (A) peptide 1-27; (A) peptide 69-80. Values are means of duplicate assays f S.D. If not represented, S.D. is smaller than symbol size.

competed for its binding to myotoxin. The competition was assessed by incubating (i) a fixed amount of heparin with varying amounts of MAb-3, and (ii) a fixed amount of MAb-3 with varying amounts of heparin. Different heparin/MAb-3 mixtures, in Iscove's medium with 1% fetal calf serum, were applied to microwells coated with myotoxin I1 (0.4 pg/well), and incubated for 4 h a t room temperature. After five washings with PBS containing 0.05% Tween 20, MAb-3 was detected with anti-mouse IgG-horseradish peroxidase (Cappel). Color was devel- oped with hydrogen peroxide and 2,2'-azino-bis-(3-ethylbenzthiazoline- 6-sulfonic acid), and recorded at 405 nm. Horse polyvalent antivenom (batch 203LQ, Instituto Clodomiro Picado) (33) was utilized as a posi- tive control for competition in the binding of MAb-3 to myotoxin 11.

Molecular Modeling Analyses-Molecular modelings were conducted on a Silicon Graphics Indigo' Extreme workstation using QUANTA/ CHARMm 3.3 and Biograf softwares (Molecular Simulations). A com- puter model of myotoxin I1 was constructed using as starting geometry the crystal structure of Agkistrodon pisciuorus pisciuorus K49 (34), available from the Brookhaven Protein Data Bank. Side chain replace- ments were made to the A. p . pisciuorus K49 structure, followed by an energy minimization of the whole molecule to an r.m.8. tolerance of the force field gradient of 0.1 kcallmo1.A using CHARMm ( E = 8, r-dependent).

A model of peptide 115-129 was also constructed by extracting its coordinates from the minimized structure of myotoxin 11. The peptide was surrounded by a water shell (outer cutoff 9 A) using Biograf. The water was first minimized to an r.m.s. tolerance of the force field gra- dient of 0.05 kcallmo1.A using CHARMm ( E = 1) while keeping the peptide fixed, whereafter the whole structure including the water was minimized again down to the same r.m.8. value. A molecular dynamics simulation was then conducted for the complex peptide-water. The sys- tem was heated to 300 K for 6 ps and equilibrated at this temperature for another 6 ps, and finally a 50-ps microcanonical simulation was run. All bond lengths were kept constant during the simulation using the SHAKE algorithm, allowing a time step of 2 fs.

RESULTS

Myotoxin ZI-The direct binding of 3H-labeled heparin to in- creasing amounts of myotoxin I1 is shown in Fig. 1. Most of the added heparin was bound by the protein readily at 5 pg of myotoxin Wassay. The ability of different unlabeled GAGS to interact with myotoxin I1 was studied by competition binding with 3H-labeled heparin, as shown in Fig. 2A and Table I. Notably, the N-desulfated N-acetylated heparin was as efficient as the unmodified heparin in the binding. Selective O-desulfa- tion of heparin resulted in a moderate decrease in the interac- tion, approximately 2-fold for 6-0-desulfated heparin, and 6-fold in the case of 2-0-desulfated heparin. The interaction

Heparin-binding Cytolytic Site of Myotoxic Phospholipase A, 29869

A 2 I O I 0.1 1 1 0 100 1000

UNLABELED POLYSACCHARIDE (pg)

c 0

’c

bp v

a ! 4

w POLYSACCHARIDE/~J MYOTOXIN II

FIG. 2. Interaction of heparins and other glycosaminoglycans with myotoxin 11. A, competitive direct binding to myotoxin I1 of 3H-labeled heparin and unlabeled GAGs. 300-1.11 incubation mixtures containing 2.5 pg of myotoxin I1 (555 nM), 1 x lo4 dpm of 3H-labeled heparin (-60 nM), and various amounts of unlabeled GAGs (1 pg of heparin, for instance, corresponding to 222 n ~ ) were passed through nitrocellulose filters as described under “Experimental Procedures.” Controls without adding unlabeled polysaccharide showed 5,200 dpm bound to myotoxin I1 and retained by the filters; blanks without myo- toxin I1 gave 60 dpm. The unlabeled polysaccharides were heparin (O), N-desulfated N-acetylated heparin (01, 6-O-desulfated heparin (A), 2-O-desulfated heparin (A), 2- and 6-O-desulfated heparin (01, human aorta heparan sulfate (B), chondroitin sulfate (01, 6-O-desulfated N- acetylated heparin ( + ), dermatan sulfate (V), and 0- and N-desulfated N-acetylated heparin (V). B, effect of heparins and other glycosamin- oglycans on the cytotoxic activity of myotoxin I1 to endothelial cells (tEnd). Myotoxin I1 (10 pg/150 pVwell), either alone or after incubation with heparins or other GAGs for 15 min at room temperature, was applied to cell cultures. Cytotoxicity was determined by the release of lactic dehydrogenase to the medium after 3 h at 37 OC, as described under “Experimental Procedures.” Lactate dehydrogenase release is expressed as a percentage, considering as 100% the activity of control cultures treated with 0.1% Triton X-100. Heparin (O), N-desulfated N-acetylated heparin (O), 6-O-desulfated heparin (A), 2-O-desulfated heparin (A), human aorta heparan sulfate (B), and chondroitin sulfate ( 0 ). The two selectively O-desulfated heparin samples were tested only at a single ratio (1 pg/pg toxin).

was even weaker with the heparin desulfated at both 2-0 and 6-0 positions, and with a low sulfated human aorta heparan sulfate. Dermatan sulfate and chondroitin sulfate competed poorly in the binding, only at concentrations about two orders of magnitude higher than that of unmodified heparin.

Neutralization of the cytolytic activity of myotoxin I1 con- formed to the binding inhibition data. Titration with the un- modified heparin indicated complete neutralization of toxin activity at a ratio of 21 pg of GAGlpg protein (Fig. 2 B ) . N- Desulfated N-acetylated heparin, 2-0- and 6-O-desulfated he- parins, all neutralized the toxin at this ratio, while no inhibi- tory effect was observed with chondroitin sulfate or low- sulfated heparan sulfate, even at GAG/protein mass ratios 10 times higher.

In direct binding of 3H-labeled oligosaccharides to myotoxin 11, the smallest heparin fragment capable of interaction was the hexasaccharide, the level of binding increasing with sac- charide length (Fig. 3A). Accordingly, tests of unlabeled oligo- saccharides for myotoxin neutralization showed that hexasac- charides provided partial protection, octasaccharides almost complete protection, and decasaccharides or larger species full protection against the cell-damaging activity of the toxin (Fig. 3B).

Macromolecular complex formation was investigated by add- ing increasing amounts of heparin to a myotoxin I1 solution, followed by measurement of the turbidity at 340 nm (Fig. 4.4). Myotoxin I1 did not absorb light at this wavelength. Maximal turbidity was obtained at a ratio of 0.2 pg of heparin/pg of protein (5 protein moleculesheparin chain), and decreased a t further additions of heparin. In the filter assay, when increas- ing amounts of 3H-labeled heparin were added to a fixed amount of myotoxin 11, the maximal retention occurred at a heparin:toxin molar ratio of l : l O , and then decreased as the proportion of heparin increased (Fig. a).

Heparin, even at very high concentrations, did not inhibit the binding of MAb-3 to myotoxin I1 (Fig. 5, A and B). As a control, a polyclonal antivenom preparation competed with MAb-3 in a dose-dependent mode.

Peptides-Of the three synthetic peptides, two (1-27 and 69-80) did not bind heparin in the filter assay (Fig. 11, although they could bind to nitrocellulose as determined by protein staining of dot-blots. In contrast, peptide 115-129 showed con- siderable heparin-binding ability, comparable to myotoxin I1 on a mass basis; on a molar basis, about 5-10 times more peptide was required to attain maximal heparin binding (Fig. 1).

Different modified heparins were tested for their ability to displace labeled native heparin from binding to peptide 115- 129. Only native unlabeled heparin showed significant displac- ing ability, all other modified heparins, heparan sulfate, chon- droitin and dermatan sulfate, being at least 100 times less potent (data not shown).

When peptide 115-129 was incubated with 3H-labeled hepa- rin oligosaccharides, the shortest oligosaccharides showing a detectable interaction were decasaccharides, the level of bind- ing increasing with increasing saccharide size (data not shown).

In the cytotoxicity system, peptide 115-129 reversed the in- hibition induced by heparin, but neither peptide 69-80 (Fig. 6) nor peptide 1-27 (not shown) did. At high concentrations, pep- tide 115-129 produced a clear cytolytic actionper se, which was enhanced if the assay was performed using PBS instead of culture medium (Fig. 7). The cytotoxic effect of peptide 115-129 was reproduced using two independently synthesized batches, and was completely abolished by preincubation with heparin (2.4 pg peptide/pg heparin) (Fig. 7). In contrast, peptides 1-27 or 69-80, at the same concentrations, did not induce any cell damage (Fig. 7).

The minimized structure of the myotoxin I1 model (Fig. 8) showed very small differences as compared with the starting A. p . pisciuorus K49 crystal structure (r.m.s. value of 0.78 A for main chain atoms). Dynamics simulations for peptide 115-129, cut out from the native protein, showed that its original backbone conformation was conserved (maximum r.m.s. devia- tion as compared with the minimized starting structure was 2.52 A).

DISCUSSION Some structural characteristics of the interaction between

heparin and myotoxin I1 were studied, and correlated to the neutralization of cytolytic activity. Myotoxin I1 bound most of the radiolabeled heparin added to the solutions, suggesting a

29870 Heparin-binding Cytolytic Site of Myotoxic Phospholipase A2

TABLE 1 Amount of GAGS required to inhibit 50% of 3H-labeled heparin binding to myotoxin II

The values (ICx,,) are calculated from Fie. 2 A .

Glycosaminoglycan

Heparin N-Desulfated N-acetylated heparin 6-0-Desulfated heparin 2-0-Desulfated heparin 2- and 6-0-desulfated heparin Human aorta heparan sulfate Chondroitin sulfate 6-0-Desulfated N-acetylated heparin Dermatan sulfate 0- and N-desulfated N-acetylated heparin

Sulfate groupddisaccharide unit Total N-Sulfates 0-Sulfates

2.5 1 1.5

1.5

1.7 1.5

1 0.7 1.8 1 1.3

0.8 1

0.6 0.3

1 0.4 0.2

1 0.7 1

0.7 1

0.3 0.3

n 60 I

4 6 8 10 1 2 14 20-24 SIZE OF OLIGOSACCHARIDES

2 4 6 8 10 1 2 14 16 18 20

SIZE OF OLIGOSACCHARIDES

FIG. 3. Interaction of heparin-derived oligosaccharides with myotoxin 11. A, direct binding of labeled oligosaccharides to myotoxin 11. -1 x lo4 dpm of3H-labeled oligosaccharides (-60 MI) were incubated with 10 pg of myotoxin I1 (2.2 p%) in 300 pl of incubation buffer. The amounts of 3H-labeled filter associated with myotoxin I1 were deter- mined by the nitrocellulose filtration procedure described under “Experimental Procedures.” Values are means of duplicate assays * S.D. and are expressed as percent of the radioactivity added to the incubation mixtures. B, neutralization of cytotoxic activity of myotoxin I1 by unlabeled oligosaccharides. Heparin oligosaccharides of different size were incubated with myotoxin I1 (15 min at room temperature) at a fixed ratio of 1 pg/pg, and then added to endothelial cell cultures. Cytotoxicity was determined by the release of lactate dehydrogenase, as in the legend for Fig. 2B. Values (mean * S.D. of duplicate assays) are expressed in relation to a control containing myotoxin I1 (10 pg/150 pvwell) alone.

non-selective mode of interaction, rather than a preferential binding to rare heparin sequences (i.e. as antithrombin; Ref. 35). The relative importance of the different sulfate groups was 2-0-sulfates > 6-0-sulfates > N-sulfates. In fact, at the O-sul- fate density present in unmodified heparin, the N-sulfates are redundant for the interaction. However, N-sulfates are essen- tial when the 0-sulfates are decreased to 50% or less of the

1%

w 0.35 0.35 0.8 2 8 35 50 70

120 >330

Potency

% of heparin 100 100 44 17 4 1 0.7 0.5 0.3 <0.1

0.060+ A

’0. *-*..-*r 0.0004 : : : : . :“ : :

*-.-.-*

MOLAR wno HEPARIN/MYOTOXIN II

0 1 2 3 4 5

O 12000 I 0

e I e

MOLAR wno HEPARIN/MYOTOXIN II

FIG. 4. Formation of macromolecular complexes between hep- arin and myotoxin II. A, turbidity at different hepardmyotoxin I1 ratios. Thirty pg of myotoxin I1 (2 w) were added to 2 ml of TBS, followed by additions of 5 pl of concentrated heparin solution to give the final indicated ratio of heparidmyotoxin 11; after each addition the mixtures were incubated for 1 min before reading the absorbance at 340 nm. B, Binding to the nitrocellulose filter a t different heparidmyotoxin I1 ratios. Increasing amounts of 3H-labeled heparin (from 500 to 200,000

buffer, to give the final molar ratios of heparidmyotoxin indicated. The dpm) were added to 5 pg of myotoxin I1 (1.1 p%) in 300 pl of incubation

amounts of [3Hlheparin filter-associated with myotoxin were deter- mined by the nitrocellulose filtration procedure described under “Ex- perimental Procedues.” Values are means of duplicate assays.

level present in the native molecule. Two selectively O-desul- fated heparins showed comparable ability to bind myotoxin, indicating that there is no specific requirement for the position of the 0-sulfates. Nevertheless, the sample retaining 2-0 sul- fate groups showed a slightly better binding ability compared to the sample having the same overall sulfation, but retaining 6-0 sulfates. This could be explained by the notion that idu- ronic acid confers supplementary conformational flexibility to the molecule (36), such that 0-sulfate groups carried by idu- ronic acid could fit better to the protein binding site(s) than 0-sulfate groups carried by glucosamine. The binding of chon- droitin sulfate and dermatan sulfate to myotoxin I1 was weaker

Heparin-binding Cytolytic Site of Myotoxic Phospholipase A, 29871

MONOCLONAL ANTIBODY DILUTION

ANTIVENOM DILUTION (reciprocal) 0-0

625 125 25 5 1

? I \

,*t \ 0, 1 0 . 3

04,y >A I 1 10 100 1000 10000

HEPARIN &/mi) 0-0

FIG. 5. Lack of competition between heparin and MAb-3 for binding to myotoxin I1 in enzyme immunoassay. A, different amounts of MAb-3 were added to myotoxin-coated microplates, either in the absence (0) or the presence (0) of a constant amount (2.5 mg/ml) of heparin. Bound MAb-3 was detected with an anti-mouse IgG-horserad- ish peroxidase conjugate. Normal mouse serum (A) was included as a control. Absorbance readings are mean 2 S.D. of triplicates. B , a con- stant amount of MAb-3 (1:20,000 dilution, selected from the curve in A ) was added to myotoxin-coated microplates, in the presence of varying amounts of heparin (0). Varying dilutions of a polyclonal horse anti- venom (0) were included as a control for competitive binding. Bound MAb-3 was detected as above.

than the binding of heparan sulfate, although the latter is less sulfated. This suggests some specificity for a heparidheparan sulfate ”backbone” in the interaction.

The minimal heparin oligosaccharide size capable of inter- acting with myotoxin 11, in both direct binding and neutraliza- tion experiments, was the hexasaccharide. Similar results were recently described in the case of human secreted class I1 P q (37).

The formation of macromolecular complexes between hepa- rin and myotoxin 11, as shown by turbidimetry, is in agreement with previous observations of precipitated complexes in agar- ose gel (101, and implies a multivalent interaction involving at least two heparin-binding sites on this protein. Since myotoxin I1 can occur as a homodimer (51, at least one binding site might be available on each monomer.

Increasing the molar ratio of heparin:myotoxin I1 beyond 1:5 disrupted the macromolecular complexes, but did not reverse the neutralizing effect. The decreased retention of the L3H]- heparin: myotoxin I1 complex on the nitrocellulose filters at high heparin proportions might be speculatively explained by a “masking“ effect of the heparin chains, interfering with the binding of protein to the filter. For comparison, the same type of experiments (conducted with a fixed amount of protein and increasing amounts of labeled heparin) gave similar results as for myotoxin 11 in the case of platelet factor 4 (data not shown;

PEPTIDE &A)

0 40 80 120 160 200

100 0-Op.115-129 /” t

8o 1 “. p.69-80

v

ii I 40 9

20

/ i /

0 0 10 20 30 40 50

PEPTIDE (pg/well)

FIG. 6. Peptide 116-129 reverses heparin neutralization of the cytolytic activity of myotoxin 11. Myotoxin I1 was mixed with hep- arin at a molar ratio of 1:1, and added to endothelial cells (10 pg toxidwell), resulting in a complete neutralization of its cytolytic activ- ity. Increasing amounts of peptides 69-80 or 115-129, at the concentra- tions indicated, were added to the heparin:toxin mixtures, incubated for 15 min at room temperature, and then assayed for cytotoxicity, to de- termine if myotoxin I1 was relieved from neutralization by heparin. Peptides alone, at the same concentrations (indicated as ~ U M in the top axis or as pglwell in the bottom axis), were devoid of cytotoxic effect (lactate dehydrogenase release or morphological changes). Each point represents mean S.D. of duplicates.

PEPTIDE (pM)

l o o t 0 100 200 300 400 500 600

I

pll5-129

v-v

w 40

20 l d 0 ” -/I - m w

0 25 50 75 100 125 150

PEPTIDE (pg/well)

FIG. 7. Cytolytic activity of peptide 116-129 on endothelial cells. Cytotoxicity assay using peptides 69-80 and 115-129 dissolved in either cell culture medium (filled symbols), or PBS (empty symbols). Cytolysis was estimated by lactate dehydrogenase release, as described under “Experimental Procedures.” Each point represents mean * S.D. of duplicates. For comparison, addition of intact myotoxin I1 to endo- thelial cells results in 100% lactate dehydrogenase release at -10 pg/ well or -5 pg/well, in culture medium or PBS, respectively. Peptide

The result obtained when peptide 115-129 was preincubated with hep- concentrations are indicated aspM (top axis) or as &well (bottom axis).

ann (2.4 pg peptide/pg heparin), in PBS, is indicated by the symbol (0).

multivalent interaction) (38), but showed classical saturation curves for antithrombin (39) and basic fibroblast growth factor (data not shown; both having monovalent heparin-binding sites).

MAb-3 has been characterized for its ability to neutralize myotoxin I1 in vivo, and in the cell model utilized in this report (10,321. Our results suggest that MAb-3 and heparin recognize different sites on the toxin, and thus, that at least two molec- ular mechanisms for the neutralization of the cytolytic activity of myotoxin I1 exist.

29872 Heparin-binding Cytolytic Site of Myotoxic Phospholipase A2

showing the region corresponding to residues 115-129 (thick cylinders) near the carboxyl terminus of the protein. B, the region corresponding to F IG. B . molecuar moael or myowxm 11, Dasea on me crysrru structure or A. p . paeczuorue m41 r q . A, myotoxm II a-carDon structure

residues 115-129 is visualized with the side chains of the amino acids. Lysines are shown in red, andhg-118 is shown in violet. Note the proximity of Lys-36 and Lys-38 to the region 115-129, enhancing its cationic character.

The epitope recognized by MAb-3 is discontinuous,2 and has not been identified. On the other hand, the heparin-binding region(s) of myotoxin I1 was searched by the use of synthetic peptides. In a study of pancreatic PLA,, evidence was obtained for a role of the 26 NH,-terminal residues in its interaction with heparin (40). The corresponding synthetic peptide (residues 1-27) of myotoxin I1 was tested for heparin binding, since the overall architecture of PLA, molecules is highly conserved (28, 41,421, but no evidence of interaction was obtained. Two syn- thetic peptides, 69-80 and 115-129, were selected for further analyses, on the basis of their high content of positively charged residues. Of these, only peptide 115-129 displayed heparin-binding ability, although with a decreased efficiency as compared to the whole protein. Interestingly, this peptide could reproduce the cytolytic effect of myotoxin 11, at higher concen- trations, while peptides 1-27 and 69-80 did not. In similarity with the action of myotoxins I1 (43) and I11 (441, the cytotoxic effect of peptide 115-129 was enhanced when the assay was performed in PBS, instead of medium. Altogether, the evidence strongly suggests that peptide 115-129 is part of a heparin- binding site of myotoxin 11, which is also involved in cytotoxic action. Consequently, the neutralizing mechanism of heparin can be explained by its binding to a cationic region of myotoxin I1 that is critical for the cytolytic mechanism.

Based on the crystal structure of A. p . pisciuorus K49 (34) and its high homology (75% sequence identity) (11) with myo- toxin 11, the three-dimensional structure of the latter was mod- eled, in order to locate peptide 115-129 and its context. The model showed that residues 115-129 are located at a highly exposed region, after the last a-helix or helix "E" (Fig. 8 A ) . This region is probably not hindered by dimerization (assuming that myotoxin I1 dimerizes similar to other PLA,s for which dimer crystal structures are known), and is clearly separated from the catalytic site (45, 46).

B. Lomonte, unpublished data.

When peptide 115-129 was subjected to a dynamics simula- tion in solution, no significant changes in its conformation could be predicted. This suggests that its lower heparin-bind- ing and cytolytic activities, in comparison to the whole protein, are probably not due to an altered conformation of the free peptide. Other possible explanations were investigated. First, the fact that myotoxin I1 can form dimers, and thus interact multivalently with heparin, would imply a higher avidity, as compared to the probable univalent interaction occurring with the small peptide. Second, visual inspection of the constructed model revealed that Lys-36 and Lys-38, although linearly dis- tant from peptide 115-129, are conformationally very close to this segment (Fig. 8B). These two lysines might contribute to the higher level of binding observed for heparin, and especially for the modified heparins and the shortest oligosaccharides, in the case of the whole protein. In addition, Lys-36 and Lys-38 could also contribute to the cytolytic mechanism, since previous studies have shown that basic PLA, myotoxins require negative surface charges on liposomes in order to disrupt them by a mechanism that is independent of an enzymatic activity (47, 48). It is interesting to note that either Lys-38, or both Lys-36 and Lys-38, together with a t least five positively charged resi- dues in the region 115-129, are conserved in several myotoxic PLA,s from Crotalid snakes (6,4942). In contrast, neurotoxic P q myotoxins from Elapid snake venoms do not exhibit these characteristics (531, suggesting possible differences in the mo- lecular mechanisms of muscle damage between neurotoxic and non-neurotoxic PLA, myotoxins.

Based on sequence comparisons, it was predicted that the cytolytic region of myotoxic PLA,s would be a cationid hydrophobic sequence located at residues 79-87, just before helix E (53, 54). No similar prediction applies to myotoxin 11. However, Lys-36 and Lys-38, together with the positively charged residues from peptide 115-129, are close to a group of three tyrosines (residues 117,119, and 120; Fig. 8B), forming a cationichydrophobic combination in the three-dimensional

Heparin-binding Cytolytic Site of Myotoxic Phospholipase A2 29873

structure. Further work is necessary to define in detail the cytolytic site, its cellular target, and its mechanism of action.

Acknowledgments-We gratefully acknowledge Drs. A. Naggi and G. Grazioli (Istituta G. Ronzoni, Milano, Italy) for the preparation of 2-0- and 6-0-desulfated heparins, Dr. W. Murphy for the preparation of human aorta heparan sulfate, and Profs. Ulf Lindahl and Ingemar Bjork for valuable advice.

REFERENCES 1. Rosenberg, P. (1990) in Handbook of Toxinology (Shier, W. T., and Mebs, D.,

2. Rosenberg, P. (1986) in Natural Toxins (Harris, J. B., ed) pp. 129-174,

4. Homsi-Brandeburgo, M. I., Queiroz, L. S., Santo-Neto, H., Rodrigues-Simioni, 3. GutiBrrez, J. M., Lomonte, B., and Cerdas, L. (1986) Toxicon 24, 885-894

5. Lomonte, B., and GutiBrrez, J. M. (1989) Toxicon 27,725-733 6. Krizaj, I., Bieber, A. L., Ritonja, A,, and Gubenoek, F. (1991) Eur: J. Biochem.

7. Kihara H., Uchikawa, R., Hattori, S., and Ohno, M. (1992) Biochem. Int. 28,

8. Chwetzoff, S., Couderc, J., Frachon, P., and Menez, A. (1989) FEES Lett. 248,

9. IQellBn, L., and Lindahl, U. (1991) Annu. Reu. Biochem. 60,443475

eds) pp. 67-277, Marcel Dekker, Inc., New York

Clarendon Press, Oxford

L., and Giglio, J. R. (1988) Toxicon 26, 615427

202, 1165-1168

895-903

1 4

10. Lomonte, B., Tarkowski, A,, Bagge, U., and Hanson, L. A. (1994) Biochem.

11. Francis, B., GutiBrrez, J. M., Lomonte, B., and Kaiser, I. I. (1991) Arch.

12. Lindahl, U., Cifonelli, J. A., Lindahl, B., and RodBn, L. (1965) J. Biol. Chem.

13. Hook, M., Riesenfeld, J., and Lindahl, U. (1982) Anal. Biochem. 119,23&245 14. Nagasawa, K., Inoue, Y., and Kamata, T. (1977) Carbohydr: Res. 68, 47-55 15. Lloyd, A. G., Embery, G., and Fowler, L. J. (1971) Biochern. Pharmacol. 20,

16. Danishefsky, I., and Steiner, H. (1965) Biochim. Biophys. Acta 101,3745 17. Jaseja, M., Rej, R. N., Sauriol, F., and Perlin, A. S. (1989) Can. J. Chem. 67,

18. Inoue, Y., and Nagasawa, K. (1976) Carbohydr: Res. 46,87-95 19. Iverius, P.-H. (1971) Biochem. J. 124,677483 20. Lycke, E., Johansson, M., Svennerholm, B., and Lindahl, U. (1991) J. Gen.

21. Shively, J. E., and Conrad, H. E. (1976) Biochemistry 16,3932-3942 22. Pejler, G., Lindahl, U., Larm, O., Scholander, E., Sandgren, E., and Lundblad,

23. Lane, D. A,, Denton, J., Flynn, A. M., Thunberg, L., and Lindahl, U. (1984)

24. Bitter, T., and Muir, H. M. (1962)Anal. Biochem. 4, 330334 25. Laemmli, U. K. (1970) Nature 227,68&685 26. Reisfeld, R. A,, Lewis, U. J., and Williams, D. E. (1962) Nature 196, 281-283 27. Valerio, R. M., Bray, A. M., Campbell, R. A,, DiPasquale, A,, Margellis, C.,

Pharmacol. 47, 1509-1518

Biochem. Biophys. 284, 352-359

240,2817-2820

637-648

1449-1456

Virol. 72, 1131-1137

A. (1988) J. Biol. Chem. 263, 5197-5201

Biochem. J. 218,725-732

Rodda, S. J., Geysen, H. M., and Maeji, N. J . (1993) Int. J. Pept. Protein Res.

28. Renetseder, R., Brunie, S., Dijkstra, B. W., Drenth, J., and Sigler, P. B. (1985) 42,l-9

J. Biol. Chem. 260, 11627-11634 29. Maccarana, M., Casu, B., and Lindahl, U. (1993) J. Biol. Chem. 268,23898-

23905 30. Bussolino, F., De Rossi, M., Sica, A., Colotta, F., Wang, J. M., Bocchietto, E.,

Padura, I. M., Bosia, A., Dejana, E., and Mantovani, A. (1991) J. Zmmunol.

31. Lomonte, B., and Kahan, L. (1988) Toxicon 26,675-689 32. Lomonte, B., GutiBrrez, J. M., Ramirez, M., and Diaz, C. (1992) Toxicon 30,

147,2122-2129

2.19-245 33. Bolafios, R., and Cerdas, L. (1980) Bol. Oficina Sanit. Panam. 88, 189-196 34. Holland, D. R., Clancy, L. L., Muchmore, S. W., Ryde, T. J., Einspahr, H. M.,

Finzel, B. C., Heinrikson, R. L., and Watenpaugh, K D. (1990) J. Biol.

35. Lindahl, U., Thunberg, L., Backstrom, G., Riesenfeld, J . , Nordling, K., and Chem. 266, 17649-17656

36. Casu, B., Petitou, M., Provasoli, M., and Sina, Y. P. (1988) Pends Biochem. Sci. Bjhrk, I. (1984) J. Biol. Chem. 269, 12368-12376

37. Dua, R., and Cho, W. (1994) Eur: J. Biochem. 221,481-490 13,221-225

38. Ibel, K., Poland, G. A,, Baldwin, J. P., Pepper, D. S., Luscombe, M., and Holbrook, J. J. (1986) Biochim. Biophys. Acta 870, 58-63

39. Maccarana, M., and Lindahl, U. (1993) Glycobiology 3,271-277 40. Diccianni, M. B., Lilly, M., McLean, L. R., Balasubramaniam, A., and

41. Westerlund, B., Nordlund, P., Uhlin, U., Eaker, D., and Eklund, H. (1992) Harmony, J. A. K. (1991) Biochemistry 30,9090-9097

42. Scott, D. L., Achari,A., Vldal, J. C., and Sigler, P. B. (1992) J. Biol. Chem. 267, FEES Lett. 301, 159-164

43. Lomonte, B., Tarkowski, A,, and Hanson, L. A. (1994) Toxicon 32,1359-1369 22645-22657

44. Bultrhn, E., Thelestam, M., and GutiBrrez, J. M. (1993) Biochim. Biophys. Acta 1179,253-259

45. Dijkstra, B. W., Kalk, K. H., Hol, W. G., and Drenth, J. (1981) J. Mol. Biol. 147, 97-123

46. Scott, D. L., White, S. P., Otwinowski, Z., Yuan, W., Gelb, M. H., and Sigler, P. B. (1990) Science 260, 1541-1546

47. Diaz, C., GutiBrrez, J . M., Lamonte, B., and GenB, J. A. (1991) Biochim. Biophys. Acta 1070, 455-460

48. Rufini, S., Cesaroni, P., Desideri, A., Farias, R., Gubensek, F., Gutikrrez, J. M., Luly, P., Massoud, R., Morero, R., and Pedersen, J. Z. (1992) Biochemistry

49. Kaiser, I. I., GutiBrrez, J . M., Plummer, D., Aird, S. D., and Odell, G. V. (1990) 31, 12424-12430

50. Yoshizumi, K., Liu, S. Y., Miyata, T., Saita, S., Ohno, M., Iwanaga, S., and Arch. Biochem. Biophys. 278,319-325

51. Liu, S. Y., Yoshizumi, K, Oda, N., Ohno, M., Tokunaga, F., Iwanaga, S., and Kihara, H. (1990) Toxicon 2 8 , 4 3 6 4

52. Maraganore, J. M., Merutka, G., Cho, W., Welches, W., Khdy, E J., and Kihara, H. (1990) J. Biochem. (Tokyo) 107,400-408

53. Kini, R. M., and Evans, H. J. (1989) Int. J. Pept. Protein Res. 34, 277-286 Heinrikson, R. L. (1984) J. Biol. Chem. 269, 13839-13843

54. Kini, R. M., and Iwanaga, S. (1986) Toxicon 24,895-905

". ~~.